We present a data structure to represent and manipulate large sets of (equal) terms (or expressions). Our initial and main motivation for this data structure is the simplification of expressions with respect to a formal theory, typically, an equational one. However, it happens that the data structure is also efficient to compute the congruence closure of a relation over a set of terms.We provide an abstract definition of the data structure, including a precise semantics, and we explain how to implement it efficiently. We prove the correctness of the proposed algorithms, with a complexity analysis and experimental results. We compare these algorithms with previous algorithms to compute the congruence closure and we also sketch how we use the data structure to tackle the expression simplification problem.
Proof-Producing Congruence Closure
